THE ARTIFICIAL NEURAL NETWORK ADAPTIVE FILTER FOR THE ULTRASONIC NONDESTRUCTIVE TESTING SIGNALS OF COARSE GRAINED MATERIALS
Liu Zhenqing, Liang Sui*, LuMingda**, Wei Moan*
*Institute of Acoustics, TongJi University, Shanghai 200092, P. R. China 1239, SIPIN Road
**Department of Electrical Engineering, TongJi University Shanghai 200092, P. R. China
Keywords: coarse grained materials, ultrasonic, NDT, artificial neural network, adaptive filter
ABSTRACT The nondestructive testing of coarse grained materials can only get low signal-to-noise ratio (SNR) signals due to the influence of grain scatter. In the past ten years, the SNR enhancement ability of the split spectrum processing (SSP) method was widely studied, and it was discovered that to some extent the SSP method was widely studied, and it was discovered that to some extent the SSP method can effectively enhance the SNR of ultrasonic flaw detection signals of coarse grained materials but with an unstable performance. [1] Adaptive filter provide a new approach to enhance the ultrasonic NDT signals of coarse grained materials. [2] In this paper, the latest developed artificial neural network (ANN) adaptive filter is used to process the ultrasonic NDT signals. Because the artificial neural network has high learning ability, the adaptive noise suppressor composed of it has a short convergent time and good suppression of grain scatter signal which with complicated waveform variation. In our experiments, the conventional adaptive is used to be compared with ANN adaptive filter. Specimens are austenitic stainless steel used in nuclear power station. Experiment results show that, ANN adaptive filter can follow up the variation of ultrasonic NDT signals smartly and it not only enhance the SNR effectively but also achieve little distortion of flaw echo. According to the distinctiveness of ultrasonic testing, we have also studied the structure of ANN adaptive filter that different from that of conventional types, which make this technique have stable performance and good suitability. Acknowledgment This work was supported by the National Natural Science Foundation P. R. China and Shanghai Municipal Education Committee for Scientific Research REFERENCES
- Bigutay N. M., et al, Analysis of Non-linear Frequency Diverse Clutters Suppression Algorithm, Ultrasonics, 1990, 28(1): 90-96
- Zhu Y. and Weight J. P., Ultrasonic Nondestructive Evaluation of Highly Scattering Materials Using Adaptive Filtering and Detection, IEEE Trans: UFFC, 1994, 41(1): 26-33
Publication Source: Trends in NDE Science & Technology; Proceedings of the 14th World Conference on Non-Destructive Testing, New Delhi, 8-13 December 1996.full paper not received
Publisher: Ashgate Publishing Company
© NDTnet
NDTnet
